Machine learning for soft and liquid molecular materials

Abstract

This review discusses three types of soft matter and liquid molecular materials, namely hydrogels, liquid crystals and gas bubbles in liquids, which are explored with an emergent machine learning approach. We summarize specific examples of the use of machine learning technique to study the structure and properties of soft matter at the molecular, microscopic and macroscopic levels. The approaches of artificial intelligence have greatly improved the prediction of material properties, stimulated the progress in modeling methodologies capable of revealing physical phenomena, and opened up new perspectives in the design and use of soft material devices. For this reason we also provide guidance on machine learning methods and recommendations on best practices for data understanding.